Please see the supplement for detailed comments.

Reliable downwelling shortwave radiation (DSR) estimation over the Tibetan Plateau (TP) is still a challenging scientific issue. This manuscript developed an improved parameterization scheme to obtain all-sky DSR based on satellite data and meteorological forcing data. The topic of the paper is interesting and it’s a good fit for the scope of ACP. The whole paper is well organized with clear logic and robust results. The description of the method is clear. Numbers of work are integrated into this paper, and abundant discussions are presented as well. However, there are still some rooms for the improvement.

Major issues:

1. The spatial resolution of DSR estimated in this paper is 1 km. The spatial resolution of DSR products for comparison is coarser than 10 km. While the scale of the stations normally represents a scale of about less than 1 km. The authors should give some explanation about their scale mismatch problem.
2. As far as I know, three atmospheric conditions (clear-sky, completely cloud-covered and partially cloud-covered) were distinguished based on cloud fraction data in previous study. In this paper, the author used MOD06 cloud product to distinguish cloud sky and clear sky conditions. Will this cause some uncertainties?
3. The derived DSR was compared with current widely used DSR products in this paper. That’s convincing. To our best knowledge, Letu et al. (2022) generated surface radiation products under all-sky conditions from the Himawari-8/AHI Next-Generation Geostationary Satellite. If the derived DSR can be compared with the latest DSR product, this paper may be more appealing. The reference is as follows,

Letu et al., A New Benchmark for Surface Radiation Products over the East Asia–Pacific Region Retrieved from the Himawari-8/AHI Next-Generation Geostationary Satellite, Bulletin of the American Meteorological Society, 103, E873-E888, 10.1175/bams-d-20-0148.1, 2022

Minor issues:

1. Figure 1: the caption is too brief. The same problems exist in other figures. Please check and modify.
2. Page 4, L116: The references cited here are too old. Are there any updated references on relevant studies?
3. L 46: “It plays a decisive role” => “It plays an important role”
4. L 55: “and other major rivers in Asia originate from the TP” => “and most major rivers in Asia originate from the TP”
5. L 57: “an important research object of global and regional energy” => “an important research object for global and regional energy”

A smart DSR retrieval scheme was developed and applied over the Tibetan Plateau. The effect of complex terrain in this region was investigated. The research looks interesting and the results look robust. My major concerns are as follows,

1. It is not clear how to derive water vapor content, whether it is derived from meteorological observations of temperature or relative humidity or directly from MODIS product.

2. It is not clear how to estimate turbidity from AOD at 550nm, my understanding is that Angstrom wavelength exponent over land is not good enough for the extrapolation.

3. It was stated by the authors the accuracy of the retrieval algorithm is highly dependent on the quality of input data, so it seems very necessary to do some sensitivity analysis on the uncertainty produced by potential errors of input products.

4. It is necessary to validate retrievals under clear and cloudy conditions, which is helpful for understanding of the uncertainty sources.

5. Measurements of DSR, especially the maintenance, data quality control, etc, should be introduced.